Nuclear relocation of the nephrin and CD2AP-bindingprotein dendrin promotes apoptosis of podocytesKatsuhiko Asanuma*, Kirk Nicholas Campbell, Kwanghee Kim, Christian Faul, and Peter Mundel†

Department of Medicine, Mount Sinai School of Medicine, New York, NY 10029

Edited by Marilyn Gist Farquhar, University of California at San Diego School of Medicine, La Jolla, CA, and approved April 28, 2007 (received for reviewFebruary 1, 2007)

Kidney podocytes and their slit diaphragms (SDs) form the finalbarrier to urinary protein loss. There is mounting evidence that SDproteins also participate in intracellular signaling pathways. The SDprotein nephrin serves as a component of a signaling complex thatdirectly links podocyte junctional integrity to actin cytoskeletaldynamics. Another SD protein, CD2-associated protein (CD2AP), isan adaptor molecule involved in podocyte homeostasis that canrepress proapoptotic TGF-� signaling in podocytes. Here we showthat dendrin, a protein originally identified in telencephalic den-drites, is a constituent of the SD complex, where it directly bindsto nephrin and CD2AP. In experimental glomerulonephritis, den-drin relocates from the SD to the nucleus of injured podocytes.High-dose, proapoptotic TGF-�1 directly promotes the nuclear importof dendrin, and nuclear dendrin enhances both staurosporine- andTGF-�1-mediated apoptosis. In summary, our results identify den-drin as an SD protein with proapoptotic signaling properties thataccumulates in the podocyte nucleus in response to glomerularinjury and provides a molecular target to tackle proteinuric kidneydiseases. Nuclear relocation of dendrin may provide a mechanismwhereby changes in SD integrity could translate into alterations ofpodocyte survival under pathological conditions.

glomerular injury � proapoptotic signaling � TGF-� signaling �slit diaphragm

G lomerular podocytes serve as the final barrier to urinaryprotein loss by the formation and maintenance of podocyte

foot processes (FP) and the interposed slit diaphragm (SD) (1).All forms of nephrotic syndrome are characterized by abnor-malities in podocytes, including retraction (effacement) of podo-cyte FP and/or molecular reorganization of the SD (1). Thediscovery of several podocyte proteins and their mutationanalysis, including nephrin (2), CD2-associated protein(CD2AP) (3), �-actinin-4 (4), podocin (5), TRPC6 (6, 7), andneph1 (8), have shed light on the pathogenesis of proteinuria andemphasized the critical role of the podocyte and the SD inmaintaining the function of the glomerular filtration barrier. Atthe cytoplasmic insertion site of the SD, ZO-1 (9), �-, �-,�-catenins (10), podocin (11), and CD2AP (3) are present. Innephrotic syndrome, the normal podocyte substructure is lost,with effacement of podocyte FP, proteinuria (1), as well asaltered phosphorylation of ZO-1 (12) and nephrin (13, 14).

Dendrin is a proline-rich protein of unknown function that wasoriginally identified in telencephalic dendrites of sleep-deprivedrats (15). In the brain, two protein variants (81 kDa, 89 kDa)were identified in cytosolic and membranous protein fractions(15). More recently, KIBRA, a WW domain-containing protein,was identified as a dendrin-interacting protein, but the functionalrelevance of this interaction remains to be established (16). Herewe show that dendrin is a component of the SD complex. Wefurther show that dendrin relocates to the nucleus of injuredpodocytes in a murine model of crescentic glomerulonephritis.Finally, we show that TGF-� promotes the nuclear translocationof dendrin and that nuclear dendrin amplifies staurosporine- andTGF-�1-induced podocyte apoptosis.

ResultsDendrin Is a Component of the SD Complex. Dendrin contains twoputative nuclear localization signals (NLSs) and three PPXYmotifs that are preserved among human, rat, and mouse (Fig.1a). To explore the renal expression of dendrin, we generatedand affinity-purified a peptide antibody against the C terminusof mouse dendrin. This antibody detected the previously de-scribed 89-kDa and 81-kDa isoforms of dendrin in the brain butonly the 81-kDa isoform in isolated glomeruli (Fig. 1b). Byimmunofluorescence microscopy of adult mouse kidney, dendrinwas detected exclusively in glomeruli (Fig. 1c), where it co-localized with the podocyte marker synaptopodin (17). Duringkidney development, dendrin was detected in the capillary loopstage of glomerular development (Fig. 1d), during which podo-cytes start developing their characteristic FP and the mature SD(18). Of note, we did not detect any nuclear localization ofdendrin during glomerular development. The precise subcellularlocalization of dendrin was determined by postembedding im-munoelectron microscopy of adult mouse kidney, and goldparticles were found at the cytoplasmic insertion of the SD intoFP (Fig. 1e).

The C Terminus of Dendrin Interacts with Nephrin. The immunogoldlabeling studies suggested an association of dendrin with the SDcomplex. Therefore, we performed GST pulldown and coimmu-noprecipitation studies to test whether dendrin interacts with thekey SD proteins nephrin, podocin, or CD2AP. When glomerularextracts were passed over GST columns, GST-dendrin, but notGST alone, specifically interacted with nephrin, CD2AP, andpodocin (Fig. 2a). On the other hand, GST-dendrin did notinteract with ZO-1, another component of the SD complex (9)(data not shown). To confirm further the interaction betweendendrin and nephrin, endogenous proteins were immunopre-cipitated from glomerular extracts of adult mice by using anti-dendrin and anti-nephrin antibodies as described before (11).Anti-dendrin precipitated dendrin and coprecipitated nephrin(Fig. 2b Left). Conversely, anti-nephrin precipitated nephrin andcoprecipitated dendrin (Fig. 2b Right). No interaction was foundwith an anti-GFP antibody serving as negative control (Fig. 2b).These results show that endogenous dendrin–nephrin complexesare present in podocytes in the kidney. Next, FLAG-nephrin wascoexpressed with GFP-tagged N- or C-terminal fragments of

Author contributions: K.A. and K.N.C. contributed equally to this work; K.A., K.N.C., C.F.,and P.M. designed research; K.A., K.N.C., K.K., and C.F. performed research; K.A., K.N.C.,K.K., C.F., and P.M. analyzed data; and K.A., K.N.C., and P.M. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Abbreviations: CD2AP, CD2-associated protein; FP, foot processes; NLS, nuclear localizationsequence; SD, slit diaphragm.

*Present address: Division of Nephrology, Department of Internal Medicine, JuntendoUniversity School of Medicine, 2-1-1 Hongo, Bunkyo-ku, Tokyo 113-8421, Japan.

†To whom correspondence should be addressed at: Division of Nephrology, Mount SinaiSchool of Medicine, One Gustave L. Levy Place, Box 1243, New York, NY 10029-6574.E-mail: peter.mundel@mssm.edu.

© 2007 by The National Academy of Sciences of the USA

10134–10139 � PNAS � June 12, 2007 � vol. 104 � no. 24 www.pnas.org�cgi�doi�10.1073�pnas.0700917104

Dow

nloa

ded

by g

uest

on

Oct

ober

14,

202

0

dendrin in HEK293 cells. As shown in Fig. 2c, the N-terminalfragment of dendrin did not coprecipitate with FLAG-nephrin.In contrast, the C-terminal fragment of dendrin was coprecipi-tated by anti-FLAG (Fig. 2c).

The C Terminus of Dendrin Interacts with CD2AP. The CD2APhomolog CIN85 interacts with dendrin (19). In podocytes,CD2AP interacts directly with nephrin (20) and podocin (11). Totest whether the observed interaction between CD2AP anddendrin (Fig. 2a) is direct or indirect through nephrin orpodocin, FLAG-dendrin and GFP-CD2AP were coexpressed inHEK293 cells, and GFP-CD2AP was coprecipitated withFLAG-dendrin (data not shown). Next, FLAG-tagged CD2APwas coexpressed with the GFP-tagged N- or C-terminal frag-ments of dendrin. Similar to nephrin, the interaction betweenCD2AP and dendrin occurred through the C terminus ofdendrin (Fig. 2d). In contrast, GFP-podocin did not coprecipi-tate with FLAG-dendrin in coimmunoprecipitation studies(data not shown).

Dendrin Binds Directly to Nephrin and CD2AP. To explore the directbinding of dendrin to nephrin and CD2AP, we conducted in vitroreconstitution studies with purified GST-dendrin, FLAG-nephrin, FLAG-CD2AP, and FLAG-podocin according to ourpreviously published protocols (21, 22). Purified GST-dendrinwas immobilized on glutathione–agarose beads and incubatedwith purified FLAG-CD2AP, FLAG-nephrin, or FLAG-

podocin. Nephrin and CD2AP bound directly to dendrin but notthe GST control (Fig. 2e). In contrast, we did not observe anybinding of dendrin to podocin (data not shown). Taken together,the C terminus of dendrin can interact directly with nephrin andCD2AP.

Nuclear Relocation of Dendrin in Experimental Glomerulonephritis.We found structural (Fig. 1) and biochemical (Fig. 2) evidencefor the association of dendrin with the SD complex under normalconditions. The presence of two putative NLSs (Fig. 1a)prompted us to test whether dendrin relocates to the podocytenucleus under pathological conditions. We analyzed the expres-sion and localization of dendrin in a murine model of glomer-ulonephritis that displays crescent formation, sclerosis, andproteinuria (23). In crescentic glomerulonephritis, injured podo-cytes form bridges between the glomerular tuft and Bowman’scapsule, and these podocyte bridges are thought to represent anearly event in the pathogenesis of crescent formation (24). Bydouble labeling deconvolution microscopy with the nuclearpodocyte marker WT-1 (25), we found an accumulation ofdendrin in the nucleus of such injured bridging podocytes at day7 after disease induction (Fig. 3 Middle). Interestingly, some ofthese bridging podocytes showed a dual subcellular distributionof dendrin both at the cell membrane and in the nucleus (Fig. 3Middle, arrow). In PBS-injected control mice, both bridgingpodocytes and nuclear expression of dendrin were absent (Fig.3 Top). In kidney sections from mice harvested on day 14 afterdisease induction, nuclear expression of dendrin was also notedin WT-1-positive nuclei of podocytes located within the cres-cents (Fig. 3 Bottom).

NLS-Mediated Nuclear Import of Dendrin. Having established anassociation of nuclear dendrin with inflammatory glomerulone-phritis, we next wanted to explore the function of dendrin in thenucleus. To this end, we first determined the intracellularlocalization of dendrin in cultured differentiated mouse podo-cytes. Nuclear localization of dendrin in differentiated culturedpodocytes was detected by double labeling confocal microscopywith DAPI (Fig. 4a Upper). We also conducted double labelingdeconvolution microscopy with �-tubulin but found no colocal-ization with microtubules (data not shown). Instead, we foundthe association of cytoplasmic dendrin with intermediate fila-

Fig. 2. The C terminus of dendrin interacts directly with nephrin and CD2AP.(a) Nephrin, CD2AP, and podocin from glomerular extracts (input) specificallyinteract with GST-dendrin but not with GST alone. (b) Coimmunoprecipitationexperiments showing that endogenous dendrin interacts with nephrin inpodocytes from isolated glomeruli. (Left) Immunoprecipitation (IP) with anti-dendrin. (Right) IP with anti-nephrin. No interaction is found with anti-GFPserving as negative control. (c) The GFP-tagged C-terminal (GFP-C-term) butnot N-terminal (GFP-N-term) fragment of dendrin coprecipitates with FLAG-nephrin but not the FLAG control. No interaction is found between FLAG-nephrin and GFP alone. (d) GFP-C-term but not GFP-N-term also interacts withFLAG-CD2AP. No interaction was found with the GFP or FLAG controls. (e)GST-dendrin but not the GST control binds directly to FLAG-nephrin andFLAG-CD2AP, respectively.

Fig. 1. Dendrin is a component of the SD complex. (a) Dendrin contains twoputative NLSs (NLS1 and NLS2), and three PPXY motifs (P1, P2, and P3). (b)Western blot analysis. In contrast to the brain, which contains both dendrinisoforms, only the 81-kDa isoform is found in isolated glomeruli. (c) Doublelabeling immunofluorescence microscopy of adult mouse kidney with synap-topodin confirms the expression of dendrin in podocytes. (d) During kidneydevelopment, dendrin is found in differentiating podocytes of the capillaryloop stage (C) but not in undifferentiated podocytes of the S-shaped bodystage (S). Arrowheads mark the kidney capsule. (e) Immunogold labeling ofultrathin frozen sections from adult mouse kidney shows the localization ofdendrin in podocyte FP at the cytoplasmic insertion site of the SD (arrows).

Asanuma et al. PNAS � June 12, 2007 � vol. 104 � no. 24 � 10135

MED

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Oct

ober

14,

202

0

ments as shown by double labeling deconvolution microscopywith vimentin (Fig. 4a Lower). Dendrin contains two putativenuclear localization sequences NLS1 and NLS2. To test thefunctional activity of these NLSs, we transfected HEK cells withGFP-tagged full-length dendrin and its deleted forms (�NLS1,�NLS2; Fig. 4b) and analyzed the subcellular distribution ofdendrin by double labeling with DAPI (Fig. 4c). Full-lengthdendrin was found in the nucleus of all transfected cells (Fig. 4cTop). In contrast, the deletion of NLS1 (�NLS1) completelyabrogated the nuclear import of dendrin (Fig. 4c Middle). Thedeletion of NLS2 did not impair the nuclear import of dendrinbecause all �NLS2-transfected cells showed nuclear localizationof GFP-dendrin (Fig. 4c Bottom). Next, the relevance of theresults in HEK cells for podocyte biology was assessed bytransfection of podocytes with the GFP-tagged dendrin con-structs (Fig. 4d). Similar to the endogenous protein (Fig. 4a),GFP-tagged full-length dendrin was found in the cytoplasm andnucleus of transfected podocytes (Fig. 4d Left). As in HEK cells,�NLS1 (Fig. 4d Center) but not �NLS2 (Fig. 4d Center) wasexcluded from the podocyte nucleus. Hence, NLS1 is necessaryand sufficient for the nuclear import of dendrin in podocytes.This finding is consistent with results by others also identifyingan essential NLS in dendrin (19).

Proapoptotic TGF-� Signaling Promotes Nuclear Import of Dendrin inPodocytes. Differentiated cultured podocytes that are main-tained in medium containing 10% FBS (26) show a dualcytoplasmic and nuclear localization of dendrin (Fig. 4a), raisingthe possibility that a signaling factor in the serum induces thenuclear import of dendrin. Consistent with this hypothesis, wefound that after serum starvation using medium containing 0.2%FBS for 24 h, dendrin was excluded from the podocyte nucleusas shown by double labeling confocal microscopy with DAPI(Fig. 5a, control). TGF-�1, which is present in significantamounts in serum, is a multifunctional cytokine that regulatesgrowth, differentiation, and apoptosis in most cells, including

podocytes (27, 28). We therefore tested whether TGF-�1 couldrescue the nuclear import of dendrin, and we found that theaddition of proapoptotic (5 ng/ml) TGF-�1 (27) to the culturemedium for 24 h restored the nuclear localization of dendrin(data not shown). To test whether TGF-� can directly promotethe nuclear import of dendrin, we analyzed the nuclear importat several time points up to 1 h after 5 ng/ml TGF-�1 treatmentby using untreated cells grown in 0.2% FBS as a control (Fig. 5a).We first detected nuclear localization of dendrin after 15 min.After 30 min, the nuclear import of dendrin was completed, andno further effect was observed after 60 min (Fig. 5a). Of note,lower concentrations of TGF-� (�1 ng/ml) did not promotenuclear translocation of dendrin (data not shown). The quanti-tative analysis of these results (n � 3) showed that 18.71 � 0.75%of control cells displayed nuclear dendrin versus 51.17 � 3.83%after 15 min (P � 0.001; t test), versus 72.67 � 1.42% after 30min (P � 0.000002; t test), versus 80.65 � 3.97% after 60 min(P � 0.0001; Fig. 5b). There was no significant differencebetween 30 and 60 min (P � not significant; t test).

Nuclear Dendrin Enhances Staurosporine and TGF-�-Mediated Apo-ptosis. TGF-� is often up-regulated in injured glomeruli (28),and TGF-� mediated podocyte apoptosis plays a prominent rolein the progression of glomerular disease (29). Now we find thatdendrin accumulates in the nucleus of injured podocytes in vivo(Fig. 3). Together with the observation that high-dose, proapop-totic TGF-� induces the nuclear import of dendrin (Fig. 5 a andb), these data raise the intriguing possibility that dendrin is amodulator of TGF-�-mediated apoptosis. To test this hypothe-sis, we transfected HEK293 cells that lack endogenous dendrin(data not shown) with GFP-dendrin, or GFP-dendrin�NLS-1,which cannot enter the nucleus. GFP alone served as negativecontrol. The transfected cells were subsequently treated withstaurosporine (1 �M, 12 h) or TGF-� (5 ng/ml, 24 h) withtransfected untreated cells serving as controls. Apoptotic cellswere detected by flow cytometry after immunolabeling with anannexin V conjugate. Transfection of GFP-tagged fusion pro-teins allowed the analysis of GFP-expressing transfected cellsonly. Full-length dendrin, capable of undergoing nuclear import,caused a 5.27 � 0.5-fold increase in staurosporine-induced

Fig. 4. NLS1-mediated nuclear import of dendrin. (a) (Upper) Confocalmicroscopy with DAPI shows that in differentiated cultured podocytes endog-enous dendrin is found in the nucleus. (Lower) Cytoplasmic dendrin is associ-ated with intermediate filaments as shown by double labeling deconvolutionmicroscopy with vimentin. (b) Schematic of GFP-dendrin, GFP-dendrin�NLS1,and GFP-dendrin�NLS2. (c) Double labeling with DAPI reveals nuclear local-ization of GFP-dendrin and GFP-dendrin�NLS2. In contrast, GFP-dendrin�NLS1 is excluded from the nucleus (arrows). (d) Similar to HEK cells,in transfected podocytes, GFP-dendrin and GFP-dendrin�NLS2 are importedinto the nucleus. In contrast, GFP-dendrin�NLS1 is excluded from the nucleusand accumulates in the cytoplasm.

Fig. 3. Nuclear translocation of dendrin in experimental glomerulonephri-tis. (Top) In PBS-injected control mice lacking bridging podocytes, dendrin isabsent from WT-1-positive podocyte nuclei. Arrowheads indicate autofluo-rescence of red blood cells. (Center) On day 7 after disease induction, dendrincolocalizes with WT-1 in the nucleus of a bridging podocyte that adheres toBowman’s capsule. Of note, there is also simultaneous labeling of the podo-cyte cell membrane (arrow). (Bottom) On day 14, in mice with active disease,dendrin can be found in the nucleus of WT-1-positive podocyte within cres-cents (arrows).

10136 � www.pnas.org�cgi�doi�10.1073�pnas.0700917104 Asanuma et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

14,

202

0

apoptosis, which was significantly greater compared with theeffect of GFP-dendrin�NLS1 (3.05 � 0.37-fold increase; P �0.02; t test) or GFP alone (3.10 � 0.26-fold increase; P � 0.03;t test) (Fig. 5c). In contrast, there was no difference betweenGFP-dendrin�NLS1 and GFP-transfected control cells (P � notsignificant; t test). The treatment of transfected HEK293 cellswith proapoptotic 5 ng/ml TGF-� yielded qualitatively similarresults, albeit as expected from previous studies (29), TGF-�1was a less potent inducer of apoptosis than staurosporine.GFP-dendrin caused 45.22 � 5.46% increase in apoptotic cells,which was significantly greater than the increase induced byGFP-dendrin�NLS1 (16.36 � 2.16%, P � 0.004) or GFP alone(11.30 � 1.78%; P � 0.002). Again, there was no significantdifference between GFP-dendrin�NLS1 and GFP-transfectedcontrol cells (P � not significant; t test). To strengthen theseresults further, we generated dendrin knockdown podocytes bylentiviral infection that showed a significant reduction of dendrinprotein expression (Fig. 5e). We repeated the apoptosis assaysand found that both staurosporine- (Fig. 5f ) and TGF-�-(Fig. 5g)induced podocyte apoptosis were significantly reduced in den-drin knockdown podocytes. For staurosporine, we found a5.34 � 0.77-fold increase for the control cells versus 2.49 � 0.27for dendrin knockdown cells (n � 3; P � 0.025, t test; Fig. 5f ).After TGF-�, treatment, control podocytes showed a 1.54 �0.11-fold increase in apoptosis versus 1.05 � 0.14 for dendrinknockdown podocytes (n � 5; P � 0.022, t test; Fig. 5g). Takentogether, nuclear dendrin enhances both staurosporine- andTGF-�-mediated apoptosis in podocytes.

DiscussionOriginally, dendrin was identified as a dendritic protein whoseexpression is differentially modulated by prolonged behavioralactivity (15). However, so far the molecular function of dendrinremained elusive. Here we identified dendrin as an enhancer ofapoptosis. The observation that dendrin, like synaptopodin (17),is found both in podocyte FP and dendritic spines adds furtherto the notion that podocytes and neurons share important

structural and functional features. Both cell types developprominent cell processes equipped with well organized micro-tubular cytoskeleton, intermediate, and actin filaments (30), andkey mechanism of process formation are shared by both celltypes (31). Similar to neurons, podocytes express the synapticvesicle molecule Rab3A and its specific effector rabphilin-3a(32). They contain structures resembling synaptic vesicles, whichcontain glutamate, and synaptotagmin 1, and undergo sponta-neous and stimulated exocytosis and recycling, with glutamaterelease (33). The idea that the SD is similar to a neurologicalsynapse is further supported by the observation that densin,originally identified as brain-specific synaptic protein of theO-sialoglycoprotein family (34), interacts in the brain with�-actinin-4 (35), the target gene of a familial form of FSGS (4)and is also found at the SD, where it interacts with nephrin (36).

In the current work we did not find any nuclear localization ofdendrin during renal development, suggesting that pathways ofglomerular injury are different from developmental pathways.The developmental expression of dendrin commences later thanthat of nephrin or CD2AP, thereby meshing well with thehypothesis that dendrin is dispensable for SD formation butrather acts as surveillance factor similar to the WT-1-interactingprotein WTIP (37) that can shuttle between adhesion junctionsand the nucleus (38). WTIP translocates from cell–cell junctionsto the nucleus in PAN-treated podocytes, where it suppressesWT-1-mediated gene expression. These results led to the hy-pothesis that WTIP monitors SD function and shuttles into thenucleus after podocyte injury (39), thereby translating changesin SD structure into altered gene expression.

The SD represents a signaling platform that contributes to theregulation of podocyte function in health and disease (40, 41).Based on recent data, it appears that at least two majorSD signaling pathways can be discerned: (i) acute, reversiblephosphotyrosine-based ZO-1 (12) and nephrin (13, 42–44)signaling leading to P-nephrin–Nck-mediated actin reorganiza-tion (14, 45). (ii) Chronic calcium (6, 7) and phosphoserine/threonine signaling-mediated regulation of cytoskeleton dynam-

Fig. 5. Nuclear dendrin promotes staurosporine- and TGF-�-mediated apoptosis. (a) Time course of TGF-�-induced nuclear import of dendrin after serumstarvation for 24 h. Nuclear translocation of dendrin can be observed after 15 min and is completed by 30 min. No further increase is seen at 60 min. (b)Quantitative analysis of TGF-�-induced nuclear import of dendrin. For details, see Results. (c and d) Nuclear (full-length) but not cytoplasmic (�NLS1) dendrinenhances staurosporine- (c) and TGF-�- (d) induced apoptosis in HEK293 cells. For details, see Results. (e) Western blot analysis reveals reduction of dendrin proteinexpression after lentiviral knockdown of dendrin (den) mRNA expression. GAPDH shows equal protein loading. ( f and g) Knockdown of dendrin expressionsignificantly reduces staurosporine- ( f) and TGF-�- (g) induced apoptosis in podocytes. For details, see Results.

Asanuma et al. PNAS � June 12, 2007 � vol. 104 � no. 24 � 10137

MED

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Oct

ober

14,

202

0

ics and podocyte survival (29, 46, 47). Both CD2AP and nephrininteract with the p85 regulatory subunit of phosphatidylinositol3-kinase in vivo, stimulating antiapoptotic Akt signaling (47).Podocytes lacking CD2AP are more susceptible to apoptosis,and mice lacking CD2AP show increased podocyte apoptosis inthe early stages of progressive glomerulosclerosis (46). Hence,CD2AP appears to have an antiapoptotic role, with apoptosisconsidered a critically important mechanism in podocyte deple-tion. In fact, elegant studies from several groups have clearlyestablished that podocyte depletion occurs in most progressiveglomerular diseases and results from podocyte loss while theremaining podocytes are unable to proliferate (48–51). In thecontext of the current results that dendrin amplifies TGF-�-induced apoptosis, it is of particular interest that podocytesundergo apoptosis at early stages in the course of progressiveglomerulosclerosis in TGF-�1-transgenic mice (29). The criticalrole of TGF-�1 in podocytes has been extended by the obser-vation that dose-dependent TGF-� signaling in podocytes can begood or bad (27). At a concentration exceeding 1 ng/ml, TGF-�signaling switches from a G0/G1 growth arrest and differentia-tion promoter to a pathophysiologic inducer of G2/M growtharrest and apoptosis in murine podocytes (27). Now, we identifynuclear, but not cytoplasmic, dendrin as a modulator of TGF-�-induced apoptosis because only high-dose TGF-�1 promotesimport of dendrin, and only nuclear dendrin increases stauro-sporine- or TGF-�1-induced apoptosis. In summary, we proposethat dendrin serves a surveillance function at the SD andrelocates to the podocyte nucleus in response to glomerularinjury, thereby providing a signaling mechanism wherebychanges in SD function modulate podocyte survival. The iden-tification of nuclear dendrin as an enhancer of staurosporine- orTGF-�-induced apoptosis defines an antisurvival signaling path-way.

MethodsAnimal Studies. Experimental glomerulonephritis was induced inadult male 129 mice aged 14–20 weeks (body weight, 28–38 g)by two consecutive i.p. injections of a sheep anti-rabbit glomer-ular antiserum (0.5 ml per 20 g of body weight per day for 2consecutive days) as described before (23). Mice (n � 3 pergroup) were killed on days 7 and 14 after the second injectionof anti-glomerular antibody, and kidneys were harvested. PBS-injected age-matched 129 mice (n � 3) served as controls.

Plasmid Constructs. A full-length cDNA clone of rat dendrin (52)was cloned in frame into a modified pGEX vector, pEGFP-C1(BD Bioscience Clontech, San Jose, CA), or pFLAG-CMV-5c(Sigma–Aldrich, St. Louis, MO). The N-terminal (amino acids1–351) and C-terminal (amino acids 352–653) fragments ofdendrin were generated by PCR and cloned into pEGFP-C1.Dendrin deletion constructs lacking NLS1 (amino acids 59–77)or NLS2 (amino acids 163–169) were generated by PCR andcloned into pEGFP-C1 and pFLAG-CMV-5a vectors. Mousenephrin cDNA (53) was cloned into pFLAG-CMV-5a. MouseCD2AP cDNA (3) was cloned into pFLAG-CMV-5a vector andpGFP-C1. All constructs were verified by DNA sequencing.

Cell Culture and Transient Transfection. Podocytes were cultured asdescribed before (26). Transient transfection of podocytes andHEK293 cells (American Type Culture Collection, Manassas,VA) was done as described previously (54). GFP fusion proteinswere analyzed by direct f luorescence microscopy in living cells orafter fixation and double labeling immunocytochemistry (54).

Generation of Polyclonal Antibodies Against Mouse Dendrin. Rabbitswere immunized with a keyhole limpet hemocyanin-conjugatedpeptide (single letter code, CEEGFIREDTRKTPQGNRERE)corresponding to the C terminus of mouse dendrin. The anti-

serum was affinity-purified with the corresponding peptidelinked to Ultralink (Pierce Chemical Co., Rockford, IL) accord-ing to the manufacturer’s instructions.

Immunofluorescence and Immunogold Electron Microscopy. Immu-nofluorescence microscopy of mouse kidney frozen sections andcultured podocytes as well as immunogold labeling were done asdescribed before (11). Double labeling of cultured podocytes andmouse kidney sections was performed with mouse monoclonalanti-vimentin and mouse monoclonal WT-1 antibodies, respec-tively. For the quantitative analysis of TGF-�-mediated nuclearimport of dendrin, pictures of 50 cells were captured by confocalmicroscopy for each time point. The data represent the mean �SEM of three independent experiments.

Western Blotting, Immunoprecipitation, and GST-Binding Assays.SDS/PAGE, Western blotting, coimmunoprecipitation fromtransfected HEK cells, and endogenous coimmunoprecipitationfrom isolated glomerular extracts were done as described pre-viously (21, 54). GST pulldown studies with GST-dendrin andglomerular protein extracts were done as described previously(21, 22). For Western blotting, the affinity-purified primaryantibody against dendrin was used at 1:200. The other primaryantibodies used in this work were as follows: rabbit anti-nephrin(53), rabbit anti-CD2AP (3), and rabbit anti-podocin (11). Tostudy the direct binding of nephrin and CD2AP to dendrin, GSTpulldown studies with purified recombinant proteins were doneas described before (21). One microgram of GST-dendrin wasimmobilized on GSH-agarose beads. Beads were washed fivetimes in 1% Triton X-100 in PBS, and 1 �g of purified FLAG-nephrin, FLAG-CD2AP, or FLAG-podocin in 500 �l of buffer(1% Triton X-100/50 mM Tris�HCl/150 mM NaCl) was added tothe beads. Reactions were incubated under rotation for 2 h at4°C, and beads were washed five times in radioimmunoprecipi-tation assay buffer (1% Triton X-100/50 mM Tris�HCl/200 mMNaCl/1 mM EDTA/1 mM EGTA/0.25% deoxycholate). Proteinswere eluted in 100 �l of sample buffer and analyzed by SDS/PAGE and immunoblotting (21).

Apoptosis Assays in HEK293 Cells. Twenty-four hours after trans-fection with GFP-dendrin, GFP-dendrin�NLS1 and GFP con-trol, apoptosis was induced in HEK cells by human TGF-�1(Roche, Mannheim, Germany) or staurosporine (Sigma–Aldrich). Cells were treated with 5 ng/ml TGF-� for 24 h or 1 �Mstaurosporine for 12 h. Transfected cells not treated with apo-ptotic stimuli served as controls. A commercial annexin V/pro-pidium iodide assay (Molecular Probes, Eugene, OR) was usedto quantify cells undergoing apoptosis. Transfected HEK293cells were harvested by rinsing with complete medium (DMEM/10% FBS/penicillin/streptomycin) and washed in cold PBS. Cellswere diluted to a concentration of 1 � 106 cells/ml in annexin-binding buffer (10 mM Hepes/140 mM NaCl/2.5 mM CaCl2, pH7.4) and incubated for 15 min with Alexa Fluor 647 annexin Vconjugate (Molecular Probes). Necrotic cells were identifiedwith propidium iodide staining. Cells were analyzed on a FACS-Calibur flow cytometer (BD Bioscience, Mountain View, CA)with data analysis of 30,000–50,000 events done on CellQuestPro (BD Bioscience).

Lentivirus Production. Two dendrin-specific oligonucleotides(GAGGCTCACCTGCTGTTGAGA, corresponding to nt 448–468 of the mouse dendrin-ORF, and GGAAGTGAGCA-GATATGGCTT, corresponding to 961–981) were cloned intothe lentiviral vector FUGW, which coexpresses GFP under theubiquitin promoter. A control FUGW vector was constructed byusing a 21-nt scrambled sequence (GACCGCGACTCGC-CGTCTGCG), which has no significant homology to any mam-

10138 � www.pnas.org�cgi�doi�10.1073�pnas.0700917104 Asanuma et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

14,

202

0

malian gene sequence. Lentivirus was produced as describedpreviously by Dittgen et al. (55).

Apoptosis Assays in Podocytes. Lentivirus-containing medium wasadded to cultured murine podocytes grown under permissiveconditions at 33°C after pretreatment with 10 �g/ml Polybrenefor 5 min. GFP expression indicating lentiviral infection ofpodocytes was noted uniformly after 48 h of transduction.Apoptotic stimuli were provided by 1 �M staurosporine for 1 hor 5 ng/ml TGF-� for 24 h. Before TGF-� treatment, podocyteswere serum-starved in 0.2% FBS-containing medium for 24 h.For the quantification of apoptotic cells, only GFP-expressinginfected podocytes were considered, where the infection effi-ciency was 70–80%.

We thank Svetlana Ratner and Mary Donnelly for excellent technicalassistance as well as Ashwini Dhume for help with confocal microscopy.We thank Stuart Shankland (University of Washington, Seattle, WA) forproviding tissue sections from mice with experimental glomerulonephri-tis. We also thank Raghu Kalluri (Beth Israel Deaconess Medical Center,Boston, MA), Peter Seeburg (MPJ for Medical Research, Heidelberg,Germany), and Andrey Shaw (Washington University, St. Louis, MO)for various cDNA constructs and antibodies, respectively. We also thankPavel Osten (Northwestern University, Chicago, IL) for providinglentiviral plasmids. This work was supported by National Institutes ofHealth Grants DK57683 and DK062472 and the George M. O’BrienKidney Center Grant DK064236 (to P.M.). K.A. was supported by aResearch Fellowship from the National Kidney Foundation, a personalgrant from Yasuhiko Tomino (Juntendo University, Tokyo, Japan), andthe Alumni Association of Juntendo University, Tokyo, Japan.

1. Somlo S, Mundel P (2000) Nat Genet 24:333–335.2. Kestila M, Lenkkeri U, Mannikko M, Lamerdin J, McCready P, Putaala H,

Ruotsalainen V, Morita T, Nissinen M, Herva R, et al. (1998) Mol Cell1:575–582.

3. Shih NY, Li J, Karpitskii V, Nguyen A, Dustin ML, Kanagawa O, Miner JH,Shaw AS (1999) Science 286:312–315.

4. Kaplan JM, Kim SH, North KN, Rennke H, Correia LA, Tong HQ, Mathis BJ,Rodriguez-Perez JC, Allen PG, Beggs AH, Pollak MR (2000) Nat Genet24:251–256.

5. Boute N, Gribouval O, Roselli S, Benessy F, Lee H, Fuchshuber A, Dahan K,Gubler MC, Niaudet P, Antignac C (2000) Nat Genet 24:349–354.

6. Reiser J, Polu KR, Moller CC, Kenlan P, Altintas MM, Wei C, Faul C, HerbertS, Villegas I, Avila-Casado C, et al. (2005) Nat Genet 37:739–744.

7. Winn MP, Conlon PJ, Lynn KL, Farrington MK, Creazzo T, Hawkins AF,Daskalakis N, Kwan SY, Ebersviller S, Burchette JL, et al. (2005) Science308:1801–1804.

8. Donoviel DB, Freed DD, Vogel H, Potter DG, Hawkins E, Barrish JP, MathurBN, Turner CA, Geske R, Montgomery CA, et al. (2001) Mol Cell Biol21:4829–4836.

9. Schnabel E, Anderson JM, Farquhar MG (1990) J Cell Biol 111:1255–1263.10. Reiser J, Kriz W, Kretzler M, Mundel P (2000) J Am Soc Nephrol 11:1–8.11. Schwarz K, Simons M, Reiser J, Saleem MA, Faul C, Kriz W, Shaw AS,

Holzman LB, Mundel P (2001) J Clin Invest 108:1621–1629.12. Kurihara H, Anderson JM, Farquhar MG (1995) Am J Physiol 268:F514–F524.13. Simons M, Schwarz K, Kriz W, Miettinen A, Reiser J, Mundel P, Holthofer H

(2001) Am J Pathol 159:1069–1077.14. Verma R, Kovari I, Soofi A, Nihalani D, Patrie K, Holzman LB (2006) J Clin

Invest 116:1346–1359.15. Neuner-Jehle M, Denizot JP, Borbely AA, Mallet J (1996) J Neurosci Res

46:138–151.16. Kremerskothen J, Plaas C, Buther K, Finger I, Veltel S, Matanis T, Liedtke T,

Barnekow A (2003) Biochem Biophys Res Commun 300:862–867.17. Mundel P, Heid HW, Mundel TM, Kruger M, Reiser J, Kriz W (1997) J Cell

Biol 139:193–204.18. Mundel P, Kriz W (1995) Anat Embryol 192:385–397.19. Kawata A, Iida J, Ikeda M, Sato Y, Mori H, Kansaku A, Sumita K, Fujiwara

N, Rokukawa C, Hamano M, et al. (2006) J Biochem (Tokyo) 139:931–939.20. Shih NY, Li J, Cotran R, Mundel P, Miner JH, Shaw AS (2001) Am J Pathol

159:2303–2308.21. Faul C, Huttelmaier S, Oh J, Hachet V, Singer RH, Mundel P (2005) J Cell Biol

169:415–424.22. Asanuma K, Yanagida-Asanuma E, Faul C, Tomino Y, Kim K, Mundel P

(2006) Nat Cell Biol 8:485–491.23. Kim YG, Alpers CE, Brugarolas J, Johnson RJ, Couser WG, Shankland SJ

(1999) Kidney Int 55:2349–2361.24. Le Hir M, Keller C, Eschmann V, Hahnel B, Hosser H, Kriz W (2001) J Am

Soc Nephrol 12:2060–2071.25. Mundlos S, Pelletier J, Darveau A, Bachmann M, Winterpacht A, Zabel B

(1993) Development (Cambridge, UK) 119:1329–1341.26. Mundel P, Reiser J, Borja AZ, Pavenstadt H, Davidson GR, Kriz W, Zeller R

(1997) Exp Cell Res 236:248–258.27. Wu DT, Bitzer M, Ju W, Mundel P, Bottinger EP (2005) J Am Soc Nephrol

16:3211–3221.

28. Bottinger EP, Bitzer M (2002) J Am Soc Nephrol 13:2600–2610.29. Schiffer M, Bitzer M, Roberts IS, Kopp JB, ten Dijke P, Mundel P, Bottinger

EP (2001) J Clin Invest 108:807–816.30. Kobayashi N, Mundel P (1998) Cell Tissue Res 291:163–174.31. Kobayashi N, Reiser J, Kriz W, Kuriyama R, Mundel P (1998) J Cell Biol

143:1961–1970.32. Rastaldi MP, Armelloni S, Berra S, Li M, Pesaresi M, Poczewski H, Langer B,

Kerjaschki D, Henger A, Blattner SM, et al. (2003) Am J Pathol 163:889–899.33. Rastaldi MP, Armelloni S, Berra S, Calvaresi N, Corbelli A, Giardino LA, Li

M, Wang GQ, Fornasieri A, Villa A, et al. (2006) FASEB J 20:976–978.34. Apperson ML, Moon IS, Kennedy MB (1996) J Neurosci 16:6839–6852.35. Walikonis RS, Oguni A, Khorosheva EM, Jeng CJ, Asuncion FJ, Kennedy MB

(2001) J Neurosci 21:423–433.36. Ahola H, Heikkila E, Astrom E, Inagaki M, Izawa I, Pavenstadt H, Kerjaschki

D, Holthofer H (2003) J Am Soc Nephrol 14:1731–1737.37. Srichai MB, Konieczkowski M, Padiyar A, Konieczkowski DJ, Mukherjee A,

Hayden PS, Kamat S, El-Meanawy MA, Khan S, Mundel P, et al. (2004) J BiolChem 279:14398–14408.

38. Nix DA, Beckerle MC (1997) J Cell Biol 138:1139–1147.39. Rico M, Mukherjee A, Konieczkowski M, Bruggeman LA, Miller RT, Khan S,

Schelling JR, Sedor JR (2005) Am J Physiol 289:F431–F441.40. Benzing T (2004) J Am Soc Nephrol 15:1382–1391.41. Huber TB, Benzing T (2005) Curr Opin Nephrol Hypertens 14:211–216.42. Palmen T, Lehtonen S, Ora A, Kerjaschki D, Antignac C, Lehtonen E,

Holthofer H (2002) J Am Soc Nephrol 13:1766–1772.43. Lahdenpera J, Kilpelainen P, Liu XL, Pikkarainen T, Reponen P, Ruotsalainen

V, Tryggvason K (2003) Kidney Int 64:404–413.44. Verma R, Wharram B, Kovari I, Kunkel R, Nihalani D, Wary KK, Wiggins RC,

Killen P, Holzman LB (2003) J Biol Chem 278:20716–20723.45. Jones N, Blasutig IM, Eremina V, Ruston JM, Bladt F, Li H, Huang H, Larose

L, Li SS, Takano T, et al. (2006) Nature 440:818–823.46. Schiffer M, Mundel P, Shaw AS, Bottinger EP (2004) J Biol Chem 279:37004–

37012.47. Huber TB, Hartleben B, Kim J, Schmidts M, Schermer B, Keil A, Egger L,

Lecha RL, Borner C, Pavenstadt H, et al. (2003) Mol Cell Biol 23:4917–4928.48. Kriz W, LeHir M (2005) Kidney Int 67:404–419.49. Kim YH, Goyal M, Kurnit D, Wharram B, Wiggins J, Holzman L, Kershaw D,

Wiggins R (2001) Kidney Int 60:957–968.50. Asano T, Niimura F, Pastan I, Fogo AB, Ichikawa I, Matsusaka T (2005) J Am

Soc Nephrol 16:2257–2262.51. Matsusaka T, Xin J, Niwa S, Kobayashi K, Akatsuka A, Hashizume H, Wang

QC, Pastan I, Fogo AB, Ichikawa I (2005) J Am Soc Nephrol 16:1013–1023.52. Herb A, Wisden W, Catania MV, Marechal D, Dresse A, Seeburg PH (1997)

Mol Cell Neurosci 8:367–374.53. Hamano Y, Grunkemeyer JA, Sudhakar A, Zeisberg M, Cosgrove D, Morello

R, Lee B, Sugimoto H, Kalluri R (2002) J Biol Chem 277:31154–31162.54. Asanuma K, Kim K, Oh J, Giardino L, Chabanis S, Faul C, Reiser J, Mundel

P (2005) J Clin Invest 115:1188–1198.55. Dittgen T, Nimmerjahn A, Komai S, Licznerski P, Waters J, Margrie TW,

Helmchen F, Denk W, Brecht M, Osten P (2004) Proc Natl Acad Sci USA101:18206–18211.

Asanuma et al. PNAS � June 12, 2007 � vol. 104 � no. 24 � 10139

MED

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Oct

ober

14,

202

0